⟡back

random.choice() vs random.choices()

Here's a little something I caught a while ago. I was expanding a Flask app that returned objects β€” vocabulary pairs β€” from the database and would display them in the template at random:

def get_random_pair():
    vocab = VocabItem.query.all() 
    random_record = random.choice(vocab) 
    random_german_word = random_record.de_word 
    english_translation = random_record.en_transl 
    number_of_records = len(vocab) 
    return render_template(
        "index.html", 
        random_german_word=random_german_word, 
        english_translation=english_translation, number_of_records=number_of_records
    )

In line 3, the queryset fetches all instances of VocabItem (via the SQLAlchemy ORM, by the way). In line 4, random_record represents one randomly-selected item from this queryset. By importing Python's in-built random module and calling the .choice() method, one non-iterable object β€” a string or an integer, say β€” will be returned in its original form.

But what if you type it wrong, like I did at first, and you don't get the result you expected? Unthinkingly, at first I used .choices() instead of .choice() in the code above. Since my method was expecting to get a non-iterable object and was instead presented with a list, it could not return the attributes that I asked for.

So what does it do?

Let's say we have a list called vegan_protein_sources:

vegan_protein_sources = ["tempeh", "tofu", "cashews", "kidney beans", "chickpeas", "oats"]

# The list serves as the 'sequence' parameter for .choices(): 
random.choices(vegan_protein_sources) 

# Output is a single, random element returned in list format: 
['tofu'] 

# Now again, but with optional parameters 'weights' and 'k': 
random.choices(
    vegan_protein_sources, 
    weights=[2, 53, 23, 12, 53, 22], 
    k=2
) 

# Output is 2 random list elements, as was specified in the 'k' parameter: 
['cashews', 'tofu']

What do you notice about the statement on line 13 in relation to the output? And what's weights, anyway?

weights is used to orchestrate the probability that these elements will be chosen. There is a list of six elements contained in the weights keyword argument β€” the same number of elements in vegan_protein_sources. So this means each element will be given that weight in the order that the integers are given in the weights parameter.

Off the top of my head, I can't really think of an exact use case in software development for .choices(), but I can imagine it comes in very handy for more scientific disciplines.

So, to summarise: .choice() for one selected element from a list, .choices() for a list of elements randomly selected from the list under certain parameters.